---
title: "Connecticut Real Estate Sales from 2011-2013"
output: 
  flexdashboard::flex_dashboard:
    source_code: embed
    theme: cosmo
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(warning = FALSE, message = FALSE)
library(flexdashboard)
library(RSocrata)
library(DT)
library(ggplot2)
library(plotly)
library(dplyr)
```


Dashboard
========================================


```{r, initial_data, include=FALSE}
url = "https://data.ct.gov/Housing-and-Development/Real-Estate-Sales-By-Town-for-2011-2012-2013/8udc-aepg"
real_estate <- read.socrata(url)

# get some small results for dashboard boxes
tbl = as.data.frame(table(real_estate$Name))

tbl$Freq <- as.numeric(tbl$Freq)

mx <- as.character(tbl$Var1[which(tbl$Freq == max(tbl$Freq))])

```


Row
-----------------------------------------------------------------------

### Number of Sold Properties

```{r, vb1}
nums <- nrow(real_estate)
valueBox(nums, icon = "fa-home")
```


### Median of Sale Price
```{r, vb2}
med <- median(real_estate$SalePrice)
valueBox(med, icon = "fa-usd")
```


### Town with Most Sold Properties
```{r, vb3}
valueBox(mx, icon = "fa-road")

```


Analysis {.storyboard}
=========================================

### **Getting the data**: Exploring open government data and developing a model to predict **sales price** based on the assesed price and name of the town using historical records from the [data.ct.gov](data.ct.gov) open data portal.

```{r, dt, echo=FALSE}

head(real_estate, n=10)
```


### **Data Cleaning & Wrangling**: The original data included punctuation for the sale and assesed price.

```{r, wrangle}

# pre-processing - remove $ punctuation
real_estate$AssessedValue <- gsub("^[[:punct:]]", "", real_estate$AssessedValue)
real_estate$SalePrice <- gsub("^[[:punct:]]", "", real_estate$SalePrice)
# change columns AssessedValue and SalePrice from chr to numeric
real_estate$AssessedValue2 <- as.numeric(real_estate$AssessedValue)
real_estate$SalePrice2 <- as.numeric(real_estate$SalePrice)
# summary numbers of reponse
summary(real_estate$SalePrice2)
# correlations
cor(real_estate$SalePrice2, real_estate$AssessedValue2) # moderate-high value

```

### **Exploratory Data Analysis**: Discovering correlations between independent and dependent variables through graphic representation.

```{r, eda, echo=FALSE}
# scatterplot 
b_real_estate <- real_estate[, - 3] # create background data for plot

s <- ggplot(real_estate, aes(x = AssessedValue2, y = SalePrice2, fill=factor(ListYear))) + 
            geom_point(data=b_real_estate, fill = "grey", alpha=0.3) +
            geom_point() +
            facet_wrap(~ ListYear) +
            theme_bw() +
            labs(title = "Real Estate Prices by Listing Year", 
                 x = "Assesed Value ($)", y = "Sale Price ($)") +
            guides(fill = FALSE)

ggplotly(s)

```


### Feature Engineering


### Modeling


Details
=========================================

*Summary of Data Source:* The Office of Policy and Management maintains a listing of all real estate sales with a sales price of $2,000 or greater that occur between October 1 and September 30 of each year. For each sale record, the file includes: town, property address, date of sale, property type (residential, apartment, commercial, industrial or vacant land), sales price, and property assessment.

**Source**: [https://data.ct.gov/Housing-and-Development/Real-Estate-Sales-By-Town-for-2011-2012-2013/8udc-aepg](https://data.ct.gov/Housing-and-Development/Real-Estate-Sales-By-Town-for-2011-2012-2013/8udc-aepg)